]> git.saurik.com Git - bison.git/blame_incremental - tests/regression.at
* data/glr.c (yytnamerr): New function.
[bison.git] / tests / regression.at
... / ...
CommitLineData
1# Bison Regressions. -*- Autotest -*-
2# Copyright (C) 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
3
4# This program is free software; you can redistribute it and/or modify
5# it under the terms of the GNU General Public License as published by
6# the Free Software Foundation; either version 2, or (at your option)
7# any later version.
8
9# This program is distributed in the hope that it will be useful,
10# but WITHOUT ANY WARRANTY; without even the implied warranty of
11# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12# GNU General Public License for more details.
13
14# You should have received a copy of the GNU General Public License
15# along with this program; if not, write to the Free Software
16# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
17# 02110-1301, USA.
18
19AT_BANNER([[Regression tests.]])
20
21
22## ------------------------- ##
23## Early token definitions. ##
24## ------------------------- ##
25
26
27AT_SETUP([Early token definitions])
28
29# Found in GCJ: they expect the tokens to be defined before the user
30# prologue, so that they can use the token definitions in it.
31
32AT_DATA_GRAMMAR([input.y],
33[[%{
34void yyerror (const char *s);
35int yylex (void);
36%}
37
38%union
39{
40 int val;
41};
42%{
43#ifndef MY_TOKEN
44# error "MY_TOKEN not defined."
45#endif
46%}
47%token MY_TOKEN
48%%
49exp: MY_TOKEN;
50%%
51]])
52
53AT_CHECK([bison -o input.c input.y])
54AT_COMPILE([input.o], [-c input.c])
55
56AT_CLEANUP
57
58
59
60## ---------------- ##
61## Braces parsing. ##
62## ---------------- ##
63
64
65AT_SETUP([Braces parsing])
66
67AT_DATA([input.y],
68[[/* Bison used to swallow the character after `}'. */
69
70%%
71exp: { tests = {{{{{{{{{{}}}}}}}}}}; };
72%%
73]])
74
75AT_CHECK([bison -v -o input.c input.y])
76
77AT_CHECK([grep 'tests = {{{{{{{{{{}}}}}}}}}};' input.c], 0, [ignore])
78
79AT_CLEANUP
80
81
82## ------------------ ##
83## Duplicate string. ##
84## ------------------ ##
85
86
87AT_SETUP([Duplicate string])
88
89AT_DATA([input.y],
90[[/* `Bison -v' used to dump core when two tokens are defined with the same
91 string, as LE and GE below. */
92
93%token NUM
94%token LE "<="
95%token GE "<="
96
97%%
98exp: '(' exp ')' | NUM ;
99%%
100]])
101
102AT_CHECK([bison -v -o input.c input.y], 0, [],
103[[input.y:6.8-14: warning: symbol `"<="' used more than once as a literal string
104]])
105
106AT_CLEANUP
107
108
109## ------------------- ##
110## Rule Line Numbers. ##
111## ------------------- ##
112
113AT_SETUP([Rule Line Numbers])
114
115AT_KEYWORDS([report])
116
117AT_DATA([input.y],
118[[%%
119expr:
120'a'
121
122{
123
124}
125
126'b'
127
128{
129
130}
131
132|
133
134
135{
136
137
138}
139
140'c'
141
142{
143
144};
145]])
146
147AT_CHECK([bison -o input.c -v input.y])
148
149# Check the contents of the report.
150AT_CHECK([cat input.output], [],
151[[Grammar
152
153 0 $accept: expr $end
154
155 1 @1: /* empty */
156
157 2 expr: 'a' @1 'b'
158
159 3 @2: /* empty */
160
161 4 expr: @2 'c'
162
163
164Terminals, with rules where they appear
165
166$end (0) 0
167'a' (97) 2
168'b' (98) 2
169'c' (99) 4
170error (256)
171
172
173Nonterminals, with rules where they appear
174
175$accept (6)
176 on left: 0
177expr (7)
178 on left: 2 4, on right: 0
179@1 (8)
180 on left: 1, on right: 2
181@2 (9)
182 on left: 3, on right: 4
183
184
185state 0
186
187 0 $accept: . expr $end
188
189 'a' shift, and go to state 1
190
191 $default reduce using rule 3 (@2)
192
193 expr go to state 2
194 @2 go to state 3
195
196
197state 1
198
199 2 expr: 'a' . @1 'b'
200
201 $default reduce using rule 1 (@1)
202
203 @1 go to state 4
204
205
206state 2
207
208 0 $accept: expr . $end
209
210 $end shift, and go to state 5
211
212
213state 3
214
215 4 expr: @2 . 'c'
216
217 'c' shift, and go to state 6
218
219
220state 4
221
222 2 expr: 'a' @1 . 'b'
223
224 'b' shift, and go to state 7
225
226
227state 5
228
229 0 $accept: expr $end .
230
231 $default accept
232
233
234state 6
235
236 4 expr: @2 'c' .
237
238 $default reduce using rule 4 (expr)
239
240
241state 7
242
243 2 expr: 'a' @1 'b' .
244
245 $default reduce using rule 2 (expr)
246]])
247
248AT_CLEANUP
249
250
251
252## ---------------------- ##
253## Mixing %token styles. ##
254## ---------------------- ##
255
256
257AT_SETUP([Mixing %token styles])
258
259# Taken from the documentation.
260AT_DATA([input.y],
261[[%token <operator> OR "||"
262%token <operator> LE 134 "<="
263%left OR "<="
264%%
265exp: ;
266%%
267]])
268
269AT_CHECK([bison -v -o input.c input.y])
270
271AT_CLEANUP
272
273
274
275## ---------------- ##
276## Invalid inputs. ##
277## ---------------- ##
278
279
280AT_SETUP([Invalid inputs])
281
282AT_DATA([input.y],
283[[%%
284?
285default: 'a' }
286%&
287%a-does-not-exist
288%-
289%{
290]])
291
292AT_CHECK([bison input.y], [1], [],
293[[input.y:2.1: invalid character: `?'
294input.y:3.14: invalid character: `}'
295input.y:4.1: invalid character: `%'
296input.y:4.2: invalid character: `&'
297input.y:5.1-17: invalid directive: `%a-does-not-exist'
298input.y:6.1: invalid character: `%'
299input.y:6.2: invalid character: `-'
300input.y:7.1-8.0: missing `%}' at end of file
301]])
302
303AT_CLEANUP
304
305
306
307## ------------------- ##
308## Token definitions. ##
309## ------------------- ##
310
311
312AT_SETUP([Token definitions])
313
314# Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
315AT_DATA_GRAMMAR([input.y],
316[%{
317#include <stdio.h>
318void yyerror (const char *s);
319int yylex (void);
320%}
321[%error-verbose
322%token MYEOF 0 "end of file"
323%token 'a' "a"
324%token B_TOKEN "b"
325%token C_TOKEN 'c'
326%token 'd' D_TOKEN
327%token SPECIAL "\\\'\?\"\a\b\f\n\r\t\v\001\377\x001\x0000ff??!"
328%%
329exp: "a" "\\\'\?\"\a\b\f\n\r\t\v\001\377\x001\x0000ff??!";
330%%
331void
332yyerror (char const *s)
333{
334 fprintf (stderr, "%s\n", s);
335}
336
337int
338yylex (void)
339{
340 return SPECIAL;
341}
342
343int
344main (void)
345{
346 return yyparse ();
347}
348]])
349
350AT_CHECK([bison -o input.c input.y])
351AT_COMPILE([input])
352AT_PARSER_CHECK([./input], 1, [],
353[syntax error, unexpected "\\'?\"\a\b\f\n\r\t\v\001\377\001\377?\?!", expecting a
354])
355AT_CLEANUP
356
357
358
359## -------------------- ##
360## Characters Escapes. ##
361## -------------------- ##
362
363
364AT_SETUP([Characters Escapes])
365
366AT_DATA_GRAMMAR([input.y],
367[%{
368void yyerror (const char *s);
369int yylex (void);
370%}
371[%%
372exp:
373 '\'' "\'"
374| '\"' "\""
375| '"' "'"
376;
377]])
378# Pacify font-lock-mode: "
379
380AT_CHECK([bison -o input.c input.y])
381AT_COMPILE([input.o], [-c input.c])
382AT_CLEANUP
383
384
385
386## -------------- ##
387## Web2c Report. ##
388## -------------- ##
389
390# The generation of the reduction was once wrong in Bison, and made it
391# miss some reductions. In the following test case, the reduction on
392# `undef_id_tok' in state 1 was missing. This is stripped down from
393# the actual web2c.y.
394
395AT_SETUP([Web2c Report])
396
397AT_KEYWORDS([report])
398
399AT_DATA([input.y],
400[[%token undef_id_tok const_id_tok
401
402%start CONST_DEC_PART
403\f
404%%
405CONST_DEC_PART:
406 CONST_DEC_LIST
407 ;
408
409CONST_DEC_LIST:
410 CONST_DEC
411 | CONST_DEC_LIST CONST_DEC
412 ;
413
414CONST_DEC:
415 { } undef_id_tok '=' const_id_tok ';'
416 ;
417%%
418]])
419
420AT_CHECK([bison -v input.y])
421AT_CHECK([cat input.output], 0,
422[[Grammar
423
424 0 $accept: CONST_DEC_PART $end
425
426 1 CONST_DEC_PART: CONST_DEC_LIST
427
428 2 CONST_DEC_LIST: CONST_DEC
429 3 | CONST_DEC_LIST CONST_DEC
430
431 4 @1: /* empty */
432
433 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';'
434
435
436Terminals, with rules where they appear
437
438$end (0) 0
439';' (59) 5
440'=' (61) 5
441error (256)
442undef_id_tok (258) 5
443const_id_tok (259) 5
444
445
446Nonterminals, with rules where they appear
447
448$accept (7)
449 on left: 0
450CONST_DEC_PART (8)
451 on left: 1, on right: 0
452CONST_DEC_LIST (9)
453 on left: 2 3, on right: 1 3
454CONST_DEC (10)
455 on left: 5, on right: 2 3
456@1 (11)
457 on left: 4, on right: 5
458
459
460state 0
461
462 0 $accept: . CONST_DEC_PART $end
463
464 $default reduce using rule 4 (@1)
465
466 CONST_DEC_PART go to state 1
467 CONST_DEC_LIST go to state 2
468 CONST_DEC go to state 3
469 @1 go to state 4
470
471
472state 1
473
474 0 $accept: CONST_DEC_PART . $end
475
476 $end shift, and go to state 5
477
478
479state 2
480
481 1 CONST_DEC_PART: CONST_DEC_LIST .
482 3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
483
484 undef_id_tok reduce using rule 4 (@1)
485 $default reduce using rule 1 (CONST_DEC_PART)
486
487 CONST_DEC go to state 6
488 @1 go to state 4
489
490
491state 3
492
493 2 CONST_DEC_LIST: CONST_DEC .
494
495 $default reduce using rule 2 (CONST_DEC_LIST)
496
497
498state 4
499
500 5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';'
501
502 undef_id_tok shift, and go to state 7
503
504
505state 5
506
507 0 $accept: CONST_DEC_PART $end .
508
509 $default accept
510
511
512state 6
513
514 3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
515
516 $default reduce using rule 3 (CONST_DEC_LIST)
517
518
519state 7
520
521 5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';'
522
523 '=' shift, and go to state 8
524
525
526state 8
527
528 5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';'
529
530 const_id_tok shift, and go to state 9
531
532
533state 9
534
535 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';'
536
537 ';' shift, and go to state 10
538
539
540state 10
541
542 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' .
543
544 $default reduce using rule 5 (CONST_DEC)
545]])
546
547AT_CLEANUP
548
549
550## --------------- ##
551## Web2c Actions. ##
552## --------------- ##
553
554# The generation of the mapping `state -> action' was once wrong in
555# extremely specific situations. web2c.y exhibits this situation.
556# Below is a stripped version of the grammar. It looks like one can
557# simplify it further, but just don't: it is tuned to exhibit a bug,
558# which disapears when applying sane grammar transformations.
559#
560# It used to be wrong on yydefact only:
561#
562# static const short int yydefact[] =
563# {
564# - 2, 0, 1, 0, 0, 2, 3, 2, 5, 4,
565# + 2, 0, 1, 0, 0, 0, 3, 2, 5, 4,
566# 0, 0
567# };
568#
569# but let's check all the tables.
570
571
572AT_SETUP([Web2c Actions])
573
574AT_KEYWORDS([report])
575
576AT_DATA([input.y],
577[[%%
578statement: struct_stat;
579struct_stat: /* empty. */ | if else;
580if: "if" "const" "then" statement;
581else: "else" statement;
582%%
583]])
584
585AT_CHECK([bison -v -o input.c input.y])
586
587# Check only the tables. We don't use --no-parser, because it is
588# still to be implemented in the experimental branch of Bison.
589[sed -n 's/ *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
590
591AT_CHECK([[cat tables.c]], 0,
592[[static const unsigned char yytranslate[] =
593{
594 0, 2, 2, 2, 2, 2, 2, 2, 2, 2,
595 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
596 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
597 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
598 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
599 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
600 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
601 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
602 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
603 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
604 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
605 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
606 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
607 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
608 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
609 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
610 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
611 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
612 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
613 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
614 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
615 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
616 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
617 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
618 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
619 2, 2, 2, 2, 2, 2, 1, 2, 3, 4,
620 5, 6
621};
622static const unsigned char yyprhs[] =
623{
624 0, 0, 3, 5, 6, 9, 14
625};
626static const yysigned_char yyrhs[] =
627{
628 8, 0, -1, 9, -1, -1, 10, 11, -1, 3,
629 4, 5, 8, -1, 6, 8, -1
630};
631static const unsigned char yyrline[] =
632{
633 0, 2, 2, 3, 3, 4, 5
634};
635static const char *const yytname[] =
636{
637 "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
638 "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
639};
640static const unsigned short int yytoknum[] =
641{
642 0, 256, 257, 258, 259, 260, 261
643};
644static const unsigned char yyr1[] =
645{
646 0, 7, 8, 9, 9, 10, 11
647};
648static const unsigned char yyr2[] =
649{
650 0, 2, 1, 0, 2, 4, 2
651};
652static const unsigned char yydefact[] =
653{
654 3, 0, 0, 2, 0, 0, 1, 3, 4, 3,
655 6, 5
656};
657static const yysigned_char yydefgoto[] =
658{
659 -1, 2, 3, 4, 8
660};
661static const yysigned_char yypact[] =
662{
663 -2, -1, 4, -8, 0, 2, -8, -2, -8, -2,
664 -8, -8
665};
666static const yysigned_char yypgoto[] =
667{
668 -8, -7, -8, -8, -8
669};
670static const unsigned char yytable[] =
671{
672 10, 1, 11, 5, 6, 0, 7, 9
673};
674static const yysigned_char yycheck[] =
675{
676 7, 3, 9, 4, 0, -1, 6, 5
677};
678static const unsigned char yystos[] =
679{
680 0, 3, 8, 9, 10, 4, 0, 6, 11, 5,
681 8, 8
682};
683]])
684
685AT_CLEANUP
686
687
688## ------------------------- ##
689## yycheck Bound Violation. ##
690## ------------------------- ##
691
692
693# _AT_DATA_DANCER_Y(BISON-OPTIONS)
694# --------------------------------
695# The following grammar, taken from Andrew Suffield's GPL'd implementation
696# of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
697# yycheck's bounds where issuing a verbose error message. Keep this test
698# so that possible bound checking compilers could check all the skeletons.
699m4_define([_AT_DATA_DANCER_Y],
700[AT_DATA_GRAMMAR([dancer.y],
701[%{
702static int yylex (AT_LALR1_CC_IF([int *], [void]));
703AT_LALR1_CC_IF([],
704[#include <stdio.h>
705static void yyerror (const char *);])
706%}
707$1
708%token ARROW INVALID NUMBER STRING DATA
709%defines
710%verbose
711%error-verbose
712/* Grammar follows */
713%%
714line: header body
715 ;
716
717header: '<' from ARROW to '>' type ':'
718 | '<' ARROW to '>' type ':'
719 | ARROW to type ':'
720 | type ':'
721 | '<' '>'
722 ;
723
724from: DATA
725 | STRING
726 | INVALID
727 ;
728
729to: DATA
730 | STRING
731 | INVALID
732 ;
733
734type: DATA
735 | STRING
736 | INVALID
737 ;
738
739body: /* empty */
740 | body member
741 ;
742
743member: STRING
744 | DATA
745 | '+' NUMBER
746 | '-' NUMBER
747 | NUMBER
748 | INVALID
749 ;
750%%
751AT_LALR1_CC_IF(
752[/* A C++ error reporting function. */
753void
754yy::parser::error (const location&, const std::string& m)
755{
756 std::cerr << m << std::endl;
757}
758
759int
760yyparse ()
761{
762 yy::parser parser;
763 parser.set_debug_level (!!YYDEBUG);
764 return parser.parse ();
765}
766],
767[static void
768yyerror (const char *s)
769{
770 fprintf (stderr, "%s\n", s);
771}])
772
773static int
774yylex (AT_LALR1_CC_IF([int *lval], [void]))
775[{
776 static int toknum = 0;
777 int tokens[] =
778 {
779 ':', -1
780 };
781 ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC. */])[
782 return tokens[toknum++];
783}]
784
785int
786main (void)
787{
788 return yyparse ();
789}
790])
791])# _AT_DATA_DANCER_Y
792
793
794# AT_CHECK_DANCER(BISON-OPTIONS)
795# ------------------------------
796# Generate the grammar, compile it, run it.
797m4_define([AT_CHECK_DANCER],
798[AT_SETUP([Dancer $1])
799AT_BISON_OPTION_PUSHDEFS([$1])
800_AT_DATA_DANCER_Y([$1])
801AT_CHECK([bison -o dancer.c dancer.y])
802AT_LALR1_CC_IF(
803 [AT_CHECK([bison -o dancer.cc dancer.y])
804 AT_COMPILE_CXX([dancer])],
805 [AT_CHECK([bison -o dancer.c dancer.y])
806 AT_COMPILE([dancer])])
807AT_PARSER_CHECK([./dancer], 1, [],
808[syntax error, unexpected ':'
809])
810AT_BISON_OPTION_POPDEFS
811AT_CLEANUP
812])
813
814AT_CHECK_DANCER()
815AT_CHECK_DANCER([%glr-parser])
816AT_CHECK_DANCER([%skeleton "lalr1.cc"])